Several major design variables were given the weight of either 1 or 0 based on (a) whether or not the study used random assignment to treatment and comparison groups, (b) whether or not the study included follow-up data, (c) whether or not the study included evidence for fidelity of treatment, (d) whether demographic differences between groups were tested and differences controlled statistically, (e) whether or not the treatment was based on a manual, and (f ) whether or not the psychometric properties of the dependent measures have been established. All of the studies used psychometrically mature instruments and were assigned a weight of 1 on this dimension. In addition to the design features of the studies, we also evaluated the extent of the information included in the research reports.